1. Field of the Invention
The present invention relates to a digital watermark embedding apparatus which is effective to prevent an illegal copy of a digital moving image signal provided through, e.g., a recording medium.
2. Description of the Related Art
With spread of an apparatus which records and reproduces digital image data such as a digital VTR or a DVD (digital versatile disk), many digital moving images which can be reproduced by such an apparatus have been provided. Further, various kinds of digital moving images are distributed through the Internet or digital television broadcasting via a broadcasting satellite, a communication satellite and others, and users have begun to be able to utilize high-quality digital moving images.
A high-quality copy of a digital moving image can be readily created on a digital signal level, and it is in danger of being limitlessly copied if any copy inhibition or copy control is not applied. Therefore, in order to avoid an illegal duplication (copy) of digital moving images or control a generation number of duplications created by regular users, there has been considered a method for adding information for a copy control to a digital moving image and preventing an illegal copy by using this additional information, thereby limiting copies.
As a technique to superimpose another additional information on a digital moving image in this manner, there is known digital watermarking. The digital watermarking is a technique to embed identification information of a copyright holder or a user of a content, right information of a copyright holder, a utilization condition of a content, confidential information required when utilizing that content, or information such as the above-described copy control information (they are referred to as watermark information) in a content such as voices, music, moving pictures, still pictures converted into digital data so that such information cannot be easily perceived, and then perform a copyright protection including a utilization control a copy control protection by detecting the watermark information from the content according to needs, or facilitate a secondary use.
As one mode of digital watermarking, there is known a mode to which a spread spectrum technique is applied. In this mode, watermark information is embedded in a digital moving image by the following procedure.
[Step E1] An image signal is multiplied by a PN (Pseudorandom Noise) series in order to perform spectrum spreading.
[Step E2] The image signal after spectrum spreading is subjected to frequency conversion (e.g., DCT conversion).
[Step E3] Watermark information is embedded by changing a value of a specific frequency component.
[Step E4] Inverse frequency conversion (e.g., IDCT conversion) is applied.
[Step E5] Spectrum inverse spreading is applied (multiplying the same PN series as that at the step E1).
On the other hand, detection of the watermark information from the digital moving image in which the watermark information is embedded in this manner is carried out by the following procedure.
[Step D1] An image signal is multiplied by a PN (Pseudorandom Noise) series (the same PN series as that at the step E1) in order to perform spectral spreading.
[Step D2] The image signal after spectrum spreading is subjected to frequency conversion (e.g., DCT conversion).
[Step D3] Attention is paid to a value of a specific frequency component, and embedded watermark information is extracted.
In order to apply digital watermarking for the purpose of preventing an illegal use, the digital watermarking must have a property (robustness) that the watermark information cannot be lost or altered by various kinds of operations or intentional attacks which are assumed to be usually applied to digital productions.
One of most effective methods of increasing the robustness is increasing an embedding intensity of the digital watermark, thereby avoiding loss of information. However, when the embedding intensity of the digital watermark is increased, a digital watermark signal is perceived as a noise, which degrades an image quality. On the other hand, when the digital watermark signal itself is visually perceivable, the confidentiality of the digital watermarking mode is weakened, which further facilitates attacks.
That is, there is a trade-off relationship between the imperceptibility of the digital watermark signal (which will be referred to as transparency) and the robustness, and increasing the robustness without deteriorating the transparency is one of the most important objects in the digital watermark embedding.